What JUCE tools can be used to implement a standalone modular polyphonic synthesizer that does not require a DAW to be fully functional as standalone audio software?
What other JUCE tools would be essential to implement standalone audio software?
In Linux would using ALSA (paired with a custom real time operating system) be a good option? https://www.alsa-project.org/wiki/Main_Page
Are Preempt-RT and or Xenomai good Linux options to implement a custom real time operating system for a modular polyphonic synthesizer?
This is just an Operating System. They used to have evaluation hardware but discontinued it.
Im not sure what their cost was but they were supporting juce when they first started.
Audio Plugins: Elk runs common audio formats (in SUSHI) such as VST2, VST3 & LV2, and the Elk Audio OS SDK allows existing software to be simply re-compiled for Elk.
I think so the raspberry pis were really expensive after covid they have tutorials here using Touch OSC which is good alternative for developing rather than getting all the hardware.
You can choose any embedded Linux board and get a JUCE Standalone application built and running on it, and then work on fine-tuning the distribution you’re using to get the best possible performance, including enabling realtime options and making configuration changes to squeeze as much performance as possible out of the distribution you’ve chosen. Its not a requirement that you do the realtime tuning up front - its also quite feasible to get the application running on a regular desktop first (if you need to bootstrap development efforts), and then test on the embedded environment.
Key to this is that you have a good build infrastructure (CI/CD) set up for your project, so that it can be simply a matter of pushing to your repo from the desktop testing environment, pulling into the embedded environment and testing.
I have experience with developing high performance realtime software on Toradex hardware, for non-audio but high-performance scientific applications, and it was quite feasible to get things running as well as possible in a workstation-Linux environment, and then deploy to the Toradex-based SOM for further testing and performance optimization.
Note that this requires a bit of dexterity when it comes to math operations and capabilities of ones compiler - but the most important workflow, imho, is having good CI/CD between the two worlds, so that very strong testing and profiling can occur as development proceeds. The Toradex SOM’s are very reliable, but a lot of this has to do with BOM costs, as well. There are a lot of modules out there - so its very important to have this CI/CD workflow ready to run when the time comes to start evaluating performance.
